COVID-19 Trends Pre- and Post-Vaccine Availability

library(tidyverse)
library(patchwork)
library(plotly)

knitr::opts_chunk$set(
  fig.width = 6,
  fig.asp = .6,
  out.width = "90%"
)

theme_set(theme_minimal() + theme(legend.position = "bottom"))

options(
  ggplot2.continuous.colour = "viridis",
  ggplot2.continuous.fill = "viridis"
)

scale_colour_discrete = scale_colour_viridis_d
scale_fill_discrete = scale_fill_viridis_d

The date 12/14/20 was used as the date when vaccines became available in the U.S. Date was converted to “day of pandemic” in order to do regression on a numeric variable instead of YYYY-MM-DD.

Data

us_covid = read_csv("./data/us.csv") %>% 
  mutate(day = 1:685,
         dc_ratio = deaths/cases)

Death/Case Ratio

total_ratio_p = 
  us_covid %>% 
  ggplot(aes(x = day, y = dc_ratio)) +
  geom_line() +
  labs(title = "U.S. Death to Cases Ratio from 1/21/20 to 12/05/21",
       x = "Day", y = "Ratio") +
  geom_vline(xintercept = 329, linetype="solid", 
                color = "red", size=0.5)

pre_vacc_ratio_p = 
us_covid %>% 
  slice(1:328) %>% 
  ggplot(aes(x = day, y = dc_ratio)) +
  geom_line() +
  labs(title = "1/21/20 to 12/13/20 (Pre-Vaccine)",
       x = "Day", y = "Ratio") +
  theme(plot.title = element_text(size=10))

post_vacc_ratio_p = 
us_covid %>% 
  slice(329:685) %>% 
  ggplot(aes(x = day, y = dc_ratio)) +
  geom_line() +
  labs(title = "12/14/20 to 12/05/21 (Post-Vaccine)",
       x = "Day", y = "Ratio") +
  theme(plot.title = element_text(size=10))

total_ratio_p / (pre_vacc_ratio_p + post_vacc_ratio_p)

Cases

total_us_case_lm = 
  lm(cases ~ day, data = us_covid) %>% 
  broom::tidy()
total_us_case_p = 
  us_covid %>% 
  ggplot(aes(x = day, y = cases)) +
  geom_point() +
  geom_smooth(method = "lm", formula = y ~ x) +
  labs(title = "U.S. COVID-19 Cases from 1/21/20 to 12/05/21",
       x = "Day", y = "Cases")

pre_vacc_case_lm = 
  us_covid %>% 
  slice(1:328) %>% 
  lm(cases ~ day, data = .) %>% 
  broom::tidy()
pre_vacc_case_p = 
  us_covid %>% 
  slice(1:328) %>% 
  ggplot(aes(x = day, y = cases)) +
  geom_point() +
  geom_smooth(method = "lm", formula = y ~ x) +
  labs(title = "1/21/20 to 12/13/20 (Pre-Vaccine)",
       x = "Day", y = "Cases") +
  theme(plot.title = element_text(size=10))

post_vacc_case_lm = 
  us_covid %>% 
  slice(329:685) %>% 
  lm(cases ~ day, data = .) %>% 
  broom::tidy()
post_vacc_case_p = 
  us_covid %>% 
  slice(329:685) %>%
  ggplot(aes(x = day, y = cases)) +
  geom_point() +
  geom_smooth(method = "lm", formula = y ~ x) +
  labs(title = "12/14/20 to 12/05/21 (Post-Vaccine)",
       x = "Day", y = "Cases") +
  theme(plot.title = element_text(size=10))

total_us_case_p / (pre_vacc_case_p + post_vacc_case_p)

Deaths

total_us_death_lm = 
  lm(deaths ~ day, data = us_covid) %>% 
  broom::tidy()
total_us_death_p = 
  us_covid %>% 
  ggplot(aes(x = day, y = deaths)) +
  geom_point() +
  geom_smooth(method = "lm", formula = y ~ x) +
  labs(title = "U.S. COVID-19 Deaths from 1/21/20 to 12/05/21",
       x = "Day", y = "Deaths")

pre_vacc_death_lm = 
  us_covid %>% 
  slice(1:328) %>% 
  lm(deaths ~ day, data = .) %>% 
  broom::tidy()
pre_vacc_death_p = 
  us_covid %>% 
  slice(1:328) %>% 
  ggplot(aes(x = day, y = deaths)) +
  geom_point() +
  geom_smooth(method = "lm", formula = y ~ x) +
  labs(title = "1/21/20 to 12/13/20 (Pre-Vaccine)",
       x = "Day", y = "Deaths") +
  theme(plot.title = element_text(size=10))

post_vacc_death_lm = 
  us_covid %>% 
  slice(329:685) %>% 
  lm(deaths ~ day, data = .) %>% 
  broom::tidy()
post_vacc_death_p = 
  us_covid %>% 
  slice(329:685) %>% 
  ggplot(aes(x = day, y = deaths)) +
  geom_point() +
  geom_smooth(method = "lm", formula = y ~ x) +
  labs(title = "12/14/20 to 12/05/21 (Post-Vaccine)",
       x = "Day", y = "Deaths") +
  theme(plot.title = element_text(size=10))

total_us_death_p / (pre_vacc_death_p + post_vacc_death_p)

Slopes

tibble(
  metric = c("intercept", "slope"),
  total_case = pull(total_us_case_lm, estimate),
  pre_vacc_case = pull(pre_vacc_case_lm, estimate),
  post_vacc_case = pull(post_vacc_case_lm, estimate),
  total_death = pull(total_us_death_lm, estimate),
  pre_vacc_death = pull(pre_vacc_death_lm, estimate),
  post_vacc_death = pull(post_vacc_death_lm, estimate)) %>% 
  knitr::kable()

metric	total_case	pre_vacc_case	post_vacc_case	total_death	pre_vacc_death	post_vacc_death
intercept	-7794675.04	-2661121.00	-3491838.72	-70361.639	-32873.6457	33469.235
slope	80715.65	42073.19	74445.31	1266.709	952.0364	1087.785

U.S. Post-vax/Pre-vax Case Ratio: 1.7694241 \(\\\) U.S. State Post-vax/Pre-vax Death Ratio: 1.1425874

states = read_csv("./data/us-states.csv") %>% 
  mutate(dc_ratio = deaths/cases)

all_states_cases = 
  states %>% 
  ggplot(aes(x = date, y = cases, color = state)) +
  geom_path() +
  labs(title = "Cumulative Cases by State",
       x = "Date",
       y = "Cases")
ggplotly(all_states_cases)

all_states_deaths = 
  states %>% 
  ggplot(aes(x = date, y = deaths, color = state)) +
  geom_path() +
  labs(title = "Cumulative Deaths by State",
       x = "Date",
       y = "Cases")
ggplotly(all_states_deaths)

all_states_dc_ratio = 
  states %>% 
  ggplot(aes(x = date, y = dc_ratio, color = state)) +
  geom_path() +
  labs(title = "Deaths/Cases Ratio by State",
       x = "Date",
       y = "Deaths/Cases Ratio")
ggplotly(all_states_dc_ratio)